Alert: an uptick in online attacks

In recent weeks AIW has documented an uptick in digital attacks against civil society activists in Azerbaijan. On February 17, AIW reported a targeted wave of attacks on political activist Abulfaz Gurbanli. Although he was able to restore access to his Facebook account, more hacking attempts were documented on February 23. At the time of writing this post, access to his Gmail has not been restored.

There were others too. At least two lawyers faced similar attacks although not as pervasive and deliberate as in the case of Gurbanli. Hackers attempted to compromise their Facebook and Telegram accounts. One lawyer said, he was receiving a barrage of phone calls and text messages from unknown numbers. Another activist, Narmin Shahmarzade said that her Facebook account was targeted too. A bug on the platform also prevented Shahmarzade’s followers from leaving comments on a post she wrote and shared in February. Since then, it was possible to fix the bug thanks to Facebook’s intervention and the post is back online. Shahmarzade originally reported that after sharing the post, her followers lost the ability to post comments. There were also attempts to hack email accounts. 

It is hard to pin down any specific reason for an uptick in digital attacks. One likely explanation is the upcoming International Women’s Day. AIW documented how last year a number of activists were targeted. Narmin Shahmarzade, said she lost access to her Facebook profile on March 9, 2021, in what looked like an attempt to discredit the activist. Shahmarzade, was among scores of women who took the streets on March 8, marking International Women’s day in the capital of Baku, and was detained by the police who prevented women from marching peacefully. In an interview with AIW, Shahmarzade said, the hacker, removed her email and changed her user name. Ahead of March 8, another activist, Gulnara Mehdiyeva was targeted by a page, that leaked her sensitive personal audio messages on Facebook. 

In 2020, AIW documented how Mehdiyeva was targeted as well. On March 8, 2020, women’s rights activist and head of the Feminist Movement of Azerbaijan Gulnara Mehdiyeva and one of the main organizers of the march, realized, someone was trying to break into her Telegram account. Then her Gmail was hacked and much of her archive including photographs and documents were “downloaded” by the attacker. In less than 48 hours Mehdiyeva’s personal Facebook account was hacked. She was removed from several Facebook groups that focus on LGBTQI and women’s rights in Azerbaijan, where she was an “admin”. Then, these groups were compromised, suspended and one was deactivated. Both groups lost thousands of subscribers and content that were shared via the Facebook group page. Next in line was Mehdiyeva’s Protonmail.

new report documents a decade of censorship in Azerbaijan

On July 16, Qurium Media Foundation released a report, “A Decade of Efforts To Keep Independent Azerbaijani Media Online”. 

The report highlights the work carried out by Qurium since 2010 assisting targeted independent and opposition online news platforms in Azerbaijan. “For more than a decade, Qurium has monitored and mitigated a wide range of cyberattacks against the websites and since 2016, no less than twenty forensics reports have been released to document our findings,” reads the new report.

Denial of Service attacks

During five years (2010-2015), Qurium mitigated dozens of denial of service attacks against Azerbaijani media, and was forced to invest in mitigation hardware and to increase its Internet capacity. Commercial mitigation of denial of service was not possible for Azeri media organizations as the average cost for such services was close to 1,000 Euro/month for a small website.

During 2014-2016, several corporate efforts made Denial of Service more difficult for the attackers, both Cloudflare (2014) and later Google (2016) started to offer free protection to journalists and human rights groups and many stress testing services (aka “booters”) since then were dismantled by FBI, such as the infamous VDOS Booter and the Mirai botnet.

After three years of research of development (2014-2017), Qurium built its own mitigation hardware and upgraded its Internet capacity by a factor of 200. Although the Denial of service attacks slowly had decreased since 2017, new challenges emerged. Internet Network Interference.

Internet Network Interference

In late 2013, a new type of challenge emerged when we discovered that websites artificially were slowed down. Instead of blocking the websites that clearly would expose the motivations and those responsible for the disruptions, the websites were slowed down by limiting the amount of bandwidth available to reach them. Qurium was forced to develop a method to detect “Internet Congestion” and to keep moving affected websites to other IP addresses to keep them online. Other large providers, such as Akamai, hosting other Azeri media was also slowed down and was unable to respond effectively to the challenge.

Exposing a coordinated cyberwar strategy

Starting from 2017, the cyberwar landscape changed. 

During that year, we received customized denial of service, pen testing and vulnerability scans and the first reports of targeted malware.

A series of diverse attacks and forensics analysis including tracing back the source of a malware sent to journalists helped us to confirm that new Ministry of Transport, Communications and High Technologies and the “hacker community” built around the government, sponsored cybersecurity events were actively targeting our hosted media.

After hosting and protecting Azeri media for almost seven years, we had no doubt about the actors behind the attacks, and could publicly document that a “State Actor” was orchestrating diverse forms of cyber attacks.

Deep Packet Inspection

Also in 2017, a new method used against independent and opposition media was identified by Qurium – the Deep Packet Inspection or shortly DPI. 

In April 2017, we identified that new technical means were implemented in several operators to block some of the websites. The Azeri authorities had invested in Deep Packet Inspection equipment to block the media outlets once and for all.

By the end of April 2017 Qurium learned that there were a court order against some of our hosted media organizations. To our surprise, the websites under Deep Packet Inspection were many more than the ones mentioned in the court order. The court order stated that the listed websites (Azadliq.info, Azadliq.org, Azerbaycansaati.com, Meydan.tv and Turan TV) were “creating threats to the legitimate interests of the state and society” and must therefore be blocked.

After two years of research between 2017-2019, Qurium identified the use of DPI hardware from Allot Communications and Sandvine inside several operators in Azerbaijan.

Website flooding, phishing, and more

By 2018, many of the “stress testing services” often used to launch the Denial of Service attacks had been dismantled world wide. The attackers were forced to find new alternatives to conduct their traffic floods aiming to take the websites offline. During another forensic investigation we traced back this new source of denial of service to Russian Fineproxy (Region40). By identifying the service provider used to conduct the attacks, we could not only expose their business practices but also their management that kindly disabled the account of the attacker.

In late 2018, Denial of Service became a second priority in the strategy to harass Azeri media and once again other means were needed.

By April 2020, Qurium could finally link the denial of service attacks launched using Fineproxy service with the very same threat actor from the Ministry of Internal Affairs: sandman. Access to sandman github account provided us with a good insight of the toolset that was being used against online media and journalists in Azerbaijan.

A final report of our findings showed even more advanced capabilities, like the ability to create fake SMS or hijack SMS sent to the journalists giving the attackers the ability to take control over their social media accounts.

Phishing remains a major attack vector against journalists and human right activists, the latest phishing campaign in early July 2021 impersonated human rights watch so as to implant a malware capable of recording the desktop and webcam or exfiltrate all important documents of the victims.

Conclusion

What started in 2010 and went on for years with Denial of service attacks using third party stress testing services was extended with more sophisticated attacks in 2017 including targeted phishing and the introduction of dedicated hardware to block the websites using technologies as DART from Allot and PCEF from Sandvine.

The national blocking of many websites, not always supported by legal court orders, has been weaponized to limit visibility of the media in the country. Despite our multiple efforts to provide alternatives to make the content available, the blocking has had a huge impact in the revenue creation of the alternative media and the growth of readership.

After the introduction of Internet blocking by means of more sophisticated deep packet inspection against alternative websites in 2018, many of the blocked media opted to increase their presence in Facebook but that has proven to be an advantageous situation for the Azeri government and their secret cyber operations as Facebook has showed a bad track record in dealing with “coordinated inauthentic behavior” in the country.

You can read the full report here.

zoom calls between senior opposition figures leaked online

Between May 13 through 17, four different video clips from private Zoom calls were leaked online. The videos were taken from calls that took place between senior members of the National Council of Democratic Forces (NCDF), an alliance representing several opposition parties in Azerbaijan.  The members of the council called the leak a cybercrime committed on behalf of the ruling government. Some have called on the authorities to investigate as this is a breach of privacy according to national legislation, while others, claimed authorities were using NSO Group’s Pegasus spyware.

Until now, no clear evidence emerged indicating that indeed, Pegasus is being used in Azerbaijan. And while AIW continues its investigation into the recent leak, here is a detailed look at other available surveillance and disruption technology the government of Azerbaijan has purchased over the recent years that have the potential of eavesdropping on users’ devices. That, combined with the recent numerous reports about the Zoom app’s security vulnerabilities may provide at least some answers.

What spyware technology Azerbaijan has purchased until now

The interest in snooping on Azerbaijani nationals is not something new for a country that has been criticized by international human rights watchdogs for years over its poor record on human rights and freedoms.

In 2012, an investigative documentary film revealed how companies owned by Teliasonera [namely Azercell in Azerbaijan at the time] “allowed for “black box” probes to be fitted with their telecommunication networks. These boxes allowed for security services and police to monitor in real-time and without any judicial oversight all communication passing through, including texts, internet traffic, and phone calls.”

Two years later, Azerbaijan investigative journalist Khadija Ismayilova revealed that the country’s largest telco had ties to the ruling family, namely to the two daughters of President Ilham Aliyev, raising questions about Internet surveillance and communications security.

The same year, Citizen Lab, identified Azerbaijan, among potential customers of Milan based Hacking Team that sold surveillance equipment called Remote Control System (RCS) to Azerbaijan as well as many other countries whose rights and freedoms record been marred with violations.

“The capabilities of its flagship product, the Remote Control System (RCS), include extracting files from a targeted device, intercepting emails and instant messaging, as well as remotely activating a device’s webcam and microphone.”

Source: New traces of Hacking Team in the wild

Among significant features of RCS are:

  • capture data that is stored on a target’s computer, even if the target never sends the information over the Internet;
  • enable government surveillance of a target’s encrypted internet communications, even when the target is connected to a network that the government cannot wiretap;

  • copy files from a computer’s hard disk, record skype calls, e-mails, instant messages, and passwords typed into a web browser;

  • turn on a device’s webcam and microphone to spy on the target

Moreover, the same CitizenLab report identified an active endpoint in Azerbaijan that was active between June and November 2013 – the year, when Azerbaijan had its presidential election [October] and accidentally announced the results of the election over an app before the voting even began.

In 2015, Organized Crime and Corruption Reporting Project (OCCRP) confirmed that the Azerbaijan government was indeed a customer of the Hacking Team. Pointing at records showing the country’s Ministry of Defense among the company’s clients.

Also in 2015, the Azerbaijan government expressed interest in purchasing Dataminr technology for its ability to “explore an individual’s past digital activity on social media and discover an individual’s interconnectivity and interactions with others on social media.”

The company’s 2015 marketing material, […] suggests that identifying individual users was a key part of Dataminr’s pitch to foreign governments by allowing users to quickly locate the “original source” behind a breaking news alert, and then find that person’s most popular tweets, what hashtags they have used in the past, and who has shared their tweets.

AIW reached out to Dataminr to confirm whether the transaction took place and received the following response:

“We currently do not have any relationship with the Government of Azerbaijan nor do we intend to do so in the future.”

The same year, the government purchased specialized security equipment – Deep Packet Inspection (DPI) to be used to monitor and block social media during the first European Games, Baku was hosting. The equipment was purchased for 3millionUSD from an Israeli company Allot Communications.

In 2016 before access to independent online news platforms is blocked, evidence shows, how the government was behind generating artificial internet network congestion within Azerbaijan to prevent access to RFERL Azerbaijan Service; VoA; and Meydan TV. The same year, first mass, spear-phishing attack targets prominent rights defender and former political prisoner Rasul Jafar.

In March 2017, the same DPI technology that purchased in 2015, is used to block some of the main independent media platforms in the country.

Also in 2017, Azerbaijan purchased another Israeli surveillance product, Verint Systems which was used in targeting of LGBTW+ on Facebook.

“I was training [clients on the use of Verint software] in Azerbaijan,” related Tal. “One day, the pupils came to me during a break and asked how they could [use the software to] determine someone’s sexual preference on Facebook. It was only later, when I read about the issue, that I discovered the country is notorious for persecuting the [LGBT] community. Suddenly things came together,” said one former Verint employee in an interview.

In general, the volume of digital attacks on representatives of civil society in Azerbaijan has been on the rise in recent years and especially since 2018. This was also highlighted in 2018 by Access Now, Digital Security Helpline. Many of these and other cases were covered here and here.

Meanwhile, AIW also looked into the possibility of Pegasus software being used in Azerbaijan following the claims made by some of the civil society representatives in the country. So far, AIW found no evidence for this to be the case. However, there is plenty of other technology available that can help the ruling government to eavesdrop and snoop around.

Taking into account Zoom vulnerabilities

Over the recent months, a number of reports on Zoom’s security vulnerabilities have also made it clear, that without E2E (end to end corruption) and with several other security-related shortcomings, Zoom does not offer, fully secure communication platform and that potential loopholes within the program may have made the leak reported in Azerbaijan possible.

  1. according to researchers at Morphisec Labs there is a Zoom app bug that can enable malicious actors to record Zoom sessions and capture chat text without any of the meeting participants’ knowledge. The malware also prevents any users in a meeting from being made aware of the recording;
  2. malicious actors can assume control of a Zoom user’s microphone or webcam;
  3. Zoom could be compelled to hand over data to governments that want to monitor online assembly or control the spread of information as activists move protests online;

The last point, is especially important, as unlike companies like Google, Facebook an Twitter, Zoom is yet to release information about whether there have been cases of government requests for data it gets, and how many of those requests it complies with. The company was encouraged to do so following an open letter and Zoom promised to publish a transparency report.

Back to Azerbaijan

Taking into account the history of surveillance and equipment purchased by government vendors over the last decade, the consistent crackdown against activists during COVID, it is likely that combined with Zoom’s security vulnerabilities, the leaked video calls were recorded by a third actor, and later leaked online for the purpose of sowing discord among opposition groups.

Opposition activist, Instagram account hacked [updated]

May 9, Azerbaijani politician, Gultekin Hajibeyli’s Instagram account hacked and taken down. Instead, a fake profile impersonating Hajibeyli was set up, with her private mobile phone number shared publicly in the profile description. Hajibeyli, was targeted online previously.

Such attacks are common in Azerbaijan, where opposition politicians and independent activists are targeted online. Account “break-ins”, impersonations, blackmailing posts, content takedown requests on YouTube for alleged copyright violations are among some of the popular harassment tactics in practice.

Unlawfully obtained personal information of intimate nature, including photos, videos, and email exchanges are commonly used to target women activists. A most recent example is an online harassment campaign launched against political activist and former political prisoner Ilkin Rustamzade’s wife, Amina Rustamzade. Leaked personal pictures were shared on Facebook and Instagram by various accounts.

On May 12, the account impersonating Hajibeyli was successfully removed from Instagram.

On May 13, a new fake Instagram profile was created.